scientific-paper-dependencies
This project gives you recommendations of which scientific papers might of interest for you based on your own bibliography. The idea standing behind the recommendation is simple. The more often a paper is cited, the more important it should be for the field of research.
The project extracts the papers of your bibliograhpy and gets their metadata (ID, referenced paper and papers citing the respective paper) via the paper's DOI from semanticscholar. Then all papers are linked and plotted, identifying which papers are mentioned often.
As the number of papers increases quickly, only the papers in your own bibliography are show and the papers which have the most links (currently the topmost 10 %). With this the graph is cleaned as much as possbile without missing to much information.
Finally, recommendations on possibly interesting papers are given. For this, the papers which are not included in your own bibliography but have the most links (currently the topmost 10 % as well) are pointed out for you to read.
Installation
The code requires Python version 3.8 or higher to run properly. I recommend to run the code in a dedicated virtual environment (for example using conda).
For installing the needed libraries, run:
pip install -r requirements.txt
Usage
All steps should be done with respect to the root directory of this repository:
- Replace the
literature.bibfile with your own file - Run
python main.py
Alternatively, you may run main.py interactively using the IPython integration in the VSCode editor.
For that, the # %% comments in main.py create "code cells" to be run one after the other.
Inspect the code in main.py further for tweaking the program to your liking.